While there is good evidence that patients with severe carpal tunnel syndrome (CTS) benefit from surgery, there is less evidence of a benefit for patients with mild to moderate disease. However, mild to moderate disease still accounts for important disability. Electrodiagnostic studies (EDS) have not been shown to accurately predict outcomes for patients with CTS. Recent advances in magnetic resonance (MR) permit high-resolution neurographic imaging of the median net-ve and pilot data suggest that MR neurographic imaging (MRNI) might be a better predictor of outcome than EDS. Thus, MRNI has the potential for playing a major role in the treatment of patients with CTS. We will test more hypotheses 1) that select patients with early, mild or moderate CTS benefit more from early surgery than with conservative therapy, and: 2) that MRNI accurately, identifies those patients more likely to benefit from early surgery. To test our hypotheses, we will recruit 500 patients with early, mild or moderate clinical CTS. After enrollment, patients will complete a standardized hand examination, EDS, symptom and functional status questionnaires, and a baseline high-resolution wrist MRNI. All patients will be treated for 2 weeks with standardized, optimal non-surgical therapy. Patients who improve within 2 weeks will be followed for I year. Patients who do not improve will be offered enrollment in trial in which they will be randomly assigned to either surgery or non-surgical therapy. We expect to randomize 200 subjects. We will offer to those patients who are randomized to non-surgical therapy an optimized, standardized hand therapy protocol, pharmaceutical and psychological intervention. We will offer localized ultrasound treatment to those who fail to improve after 3 months and surgery to patients who still have failed to improve after 6 months. The primary outcome measure will be the Carpal Tunnel Syndrome Assessment Questionnaire (CTSAQ) functional status index at 12 months. Our study design will allow us to determine if there are patients who benefit from early carpal runnel release and whether MRNI accurately identifies them.